So What is SIP?
SIP (Session Initiation Protocol) is a TCP/IP Application Layer control protocol designed for the creation of voice and/or multimedia sessions across an internetwork. The sessions can comprise voice (telephony), multimedia (voice and video), distributed games and other applications. Both point to point and conferencing is possible with SIP through the use of either unicast or multicast sessions. The SIP terminals or endpoints as they are known use addresses similar to email addresses or Web URLs, a typical example being email@example.com or firstname.lastname@example.org, where the element before the @ sign represents the device and to the left of the @ represents the domain to which the device belongs. Devices such as telephones register with a SIP server to become part of the telephone system in much the same way as traditional telephones would be registered with a local PABX (Private Automatic Branch Exchange). In fact we often refer to a SIP Server providing a telephone services as an IP PBX (Internet Protocol PBX).
SIP is used in conjunction with two other TCP/IP protocols, RTP (Real Time Protocol) and RTCP (Real Time Transport Control Protocol), which are used to transport the digitised media samples and provide a reporting mechanism to report on quality and also to provide inter-media synchronization when multimedia sessions contain both voice and video. You can think of SIP as being the equivalent of a telecommunications signalling protocol and RTP / RTCP as the means of transporting the telephone information.
SIP Servers can perform a number of roles:
- Registrar Server – Used to register SIP endpoints and verify their authenticity.
- Proxy Server – Acting as the middleman to connect endpoints together, but to allow the endpoints to transfer media directly.
- Redirection Server – Providing users with the ability to be mobile and still send and receive calls.
Traditionally, the service provider or carrier will port a subscriber's number(s) into its own exchange network or hosting platform, which will be located in a secure data centre. These numbers are then presented across a private IP network Internet connection back to the subscriber's premises.
The first approach to this used analog frequency division multiplexing in an approach known as carrier telephony. The TDM Trunk allowed for an individual telephone conversation across a single 64Kbps channel, with the customer and service provider predicting how many simultaneous calls would be made across the trunk at the “Busy Hour”. This determined how many channels were needed on the trunk, and trunks were normally comprised of full or partial E1 or T1 primary rate ISDN lines.
TDM trunk lines are being replaced by SIP Trunking which can also use E1 or T1 lines, but un-channelized which means it is just a single IP connection. The amount of bandwidth required on the trunk is once again determined by the maximum number of simultaneous calls expected at the busy hour, but in this case the bandwidth requirement is determined by the codec in use. Typically 96Kbps per call for G.711 codec and 28Kbps per call when G.729 codec is in use.
How much do SIP Trunks cost my business?
A lot of service providers charge for channels on a SIP Trunk in much the same way they charged for TDM trunks, by billing per minute per channel on the trunk plus a setup fee. Because of the competition for customers the cost is being driven down and a lot of service providers are now charging a flat fee per trunk depending on the amount of bandwidth required.
This article was written by David Christie, MD at NSTUK Ltd who specialise in the delivery of Instructor-Led training courses including Introduction to Next Generation Networks training courses. Visit www.nstuk.com/scheduled-courses.html